-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathHomework2Problem5.py
81 lines (65 loc) · 2.77 KB
/
Homework2Problem5.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
import math
import random
import numpy as np
def bgd_l2(data, y, w, eta, delta, lam, num_iter):
new_w = []
history_fw = []
new_w = w
for i in range(num_iter):
w_g = 0
length = len(y)
history = 0
reg = 0
for j in range(0, length):
if y[j] >= (np.dot(np.transpose(new_w), data[j]) + delta):
w_g += -(2/length) * data[j] * (y[j] - np.dot(np.transpose(new_w), data[j]) - delta) + (2 * lam * new_w)
elif y[j] <= (np.dot(np.transpose(new_w), data[j]) - delta):
w_g += -(2/length) * data[j] * (y[j] - np.dot(np.transpose(new_w), data[j]) + delta) + (2 * lam * new_w)
else:
w_g = (2 * lam * new_w)
for j in range(0, length):
if y[j] >= (np.dot(np.transpose(new_w), data[j]) + delta):
history += ((y[j] - np.dot(np.transpose(new_w), data[j]) - delta)**2)
elif y[j] <= (np.dot(np.transpose(new_w), data[j]) - delta):
history += ((y[j] - np.dot(np.transpose(new_w), data[j]) + delta)**2)
new_w -= eta * w_g
for j in range(0, len(data[j])):
reg += w[j]**2
reg *= lam
history *= (1 / length)
history += reg
history_fw.append(history)
return new_w, history_fw
def sgd_l2(data, y, w, eta, delta, lam, num_iter, i=-1):
new_w = []
history_fw = []
new_w = w
if i == -1:
end = num_iter
else:
end = 2
for i in range(1, end):
w_g = 0
length = len(y)
history = 0
reg = 0
randomIndex = random.randint(0, length-1)
if y[randomIndex] >= (np.dot(np.transpose(new_w), data[randomIndex]) + delta):
w_g += -(2/length) * data[randomIndex] * (y[randomIndex] - np.dot(np.transpose(new_w), data[randomIndex]) - delta) + (2 * lam * new_w)
elif y[randomIndex] <= (np.dot(np.transpose(new_w), data[randomIndex]) - delta):
w_g += -(2/length) * data[randomIndex] * (y[randomIndex] - np.dot(np.transpose(new_w), data[randomIndex]) + delta) + (2 * lam * new_w)
else:
w_g = (2 * lam * new_w)
for j in range(0, length):
if y[j] >= (np.dot(np.transpose(new_w), data[j]) + delta):
history += ((y[j] - np.dot(np.transpose(new_w), data[j]) - delta)**2)
elif y[j] <= (np.dot(np.transpose(new_w), data[j]) - delta):
history += ((y[j] - np.dot(np.transpose(new_w), data[j]) + delta)**2)
new_w -= ((eta / i**(1/2)) * w_g)
for j in range(0, len(data[j])):
reg += w[j]**2
reg *= lam
history *= (1 / length)
history += reg
history_fw.append(history)
return new_w, history_fw